A little-known technique exists wherein one may measure directly the phase of combinations of reflections. This becomes possible when a crystal is oriented in an x-ray beam so that two rays (or waves) are excited simultaneously, that is, there are two reciprocal-latticet points on the Ewald sphere at once. When this is true, as we'll develop below, there will be an interaction between the second reflection g and the first reflection h via a third reflection h-g; each diffracted beam becomes the source for at least one more reflection. The result of the interactions among these three beams is that the phase and amplitude of each diffracted ray is modulated in a way that can be interpreted in terms of the phase of a "triplet" of reflections, that is, the sum of three phases. The indices of the triplet have the relationship that h, + h2 + h3 = 0. In this case h, = g, h2 = h-g, and h3 = -h. Lipscomb, for example, proposed early that one might measure relative phases of diffraction maxima by multiple reflection'. This was exploited by others, for example Hart and Lang', but without turning it into a useful technique. More recently, Post proposed that one should exploit synchrotron radiation to apply the method to imperfect crystals'. Quite recently, Weckert and HUmmer have reduced the method to practice for a range of crystals, including macromoleculeS4. It will pay to describe the physics behind the effect in a general way, to describe how it is applied, and to show a little about results that can be obtained. (Much of this description comes from Weckert and Hummer4.)